Accumulation of 2-C-Methyl-d-erythritol as a Major Osmoticum in Petals During Flower Opening in Phlox drummondii

Abstract

Flower opening is associated with the expansion of petal cells. To understand the role played by soluble carbohydrates during cell expansion associated with petal growth, changes in soluble carbohydrate concentrations in petal limbs during flower opening in Phlox drummondii were investigated. The size of adaxial and abaxial epidermal cells in the petal limbs gradually increased during flower opening. 2-C-Methyl-d-erythritol (2-C-ME) was identified using ¹H-NMR in P. drummondii petals. 2-C-ME was the most abundant carbohydrate in the petal limbs at five developmental stages, with the concentration of glucose the second highest, although the concentration of the latter was half of that of the 2-C-ME concentration in all five stages. The concentrations of 2-C-ME and glucose increased during flower opening. In contrast, inorganic ion concentrations did not increase during flower opening. The osmotic potential of petal limbs decreased considerably during the final stage of flower opening; this decrease could in part be attributed to the increasing 2-C-ME concentration. Transmission electron microscopic observations showed that the petal limb cells in open flowers were occupied primarily by the vacuole. The concentration of 2-C-ME in the vacuole was estimated to be 131 mM, which was much higher than the concentrations of the other carbohydrates. We conclude that the accumulation of 2-C-ME in the vacuole at a high concentration acts as an osmoticum, decreasing the osmotic potential of petal limbs and thereby increasing turgor pressure, which is thought to be involved in cell expansion of petal limbs during flower opening.

Introduction

The process of flower bud development and subsequent flower opening is associated with petal growth. In Gaillardia grandiflora (Koning, 1984), cell division in petals ceased at an early stage of flower opening. Similarly, petal growth during flower opening has been shown to occur mainly as a result of cell expansion in rose (Yamada et al., 2009b), Tweedia caerulea (Norikoshi et al., 2013), and Eustoma grandiflorum (Norikoshi et al., 2016). These findings indicate that the petal growth associated with flower opening depends on cell expansion.

In many plant species, including rose (Yamada et al., 2009a), carnation (Ichimura et al., 1998), and chrysanthemum (Ichimura et al., 2000b), glucose and fructose contents in the petals increase during flower opening. In rose petals, the osmotic potential of the symplast decreases during flower opening, and this decrease is mainly attributed to increases in sugar concentrations (Yamada et al., 2009a). Moreover, the addition of sugars, including glucose, fructose, and sucrose, to vase water promotes opening of cut flowers in many plants (Ichimura et al., 2000a, 2003). These findings suggest that the accumulation of soluble carbohydrates decreases the osmotic potential of petal cells. This decrease may facilitate water influx to cells, thereby increasing turgor pressure, which is thought to be involved in cell expansion.

Inorganic ions are known to contribute to the regulation of osmotic balance in plants (Beauzamy et al., 2014). They are key compounds involved in the regulation of osmotic potential in vegetative tissues (Cram, 1976; Leigh and Tomos, 1993). In rose petals, inorganic ions play a major role as osmotica at the bud stage, but their concentrations do not increase during flower opening (Yamada et al., 2009a).

For cell growth in some plant tissues, including the fruits of apple (Yamaki and Ino, 1992), melon (Ofosu-Anim and Yamaki, 1994), and grape berry (Keller and Shrestha, 2014), sugars accumulate in the cell vacuole. In the petals of tulip (Wagner, 1979) and rose (Yamada et al., 2009a), glucose and fructose accumulate mainly in the vacuole. In E. grandiflorum, in addition to glucose, sucrose accumulates in the vacuole of petal cells (Norikoshi et al., 2016). To maintain osmotic balance with the vacuole, some soluble carbohydrates, including mannitol and pinitol, accumulate in the cytoplasm (Paul and Cockburn, 1989; Norikoshi et al., 2015). In addition, soluble sugars accumulate in the apoplast in some sink organs, such as tomato fruits (Damon et al., 1988) and sugarcane stems (Welbaum and Meinzer, 1990). In rose petals, glucose and fructose accumulate in the apoplast during flower opening (Yamada et al., 2009a). Similarly, glucose accumulates in the apoplast in E. grandiflorum petals (Norikoshi et al., 2016). Thus, the determination of subcellular carbohydrate concentrations can contribute to our understanding of the mechanisms underlying cell expansion associated with petal growth.

Phlox spp. are important ornamental plants that are used as both cut flowers and potted plants. 2-C-Methyl-d-erythritol (2-C-ME) has been shown to be the major soluble carbohydrate in the petals of Phlox subulata, a common bedding plant (Enomoto et al., 2004), suggesting that 2-C-ME may be an important compound in regulating osmotic potential in P. subulata petals. P. subulata exhibits annual flowering that occurs during spring under natural environmental conditions. In contrast, the flowering of P. drummondii occurs without specific environmental regulation when the plants reach their adult phase. Thus, P. drummondii appears to be a suitable choice for conducting basic research.

The nonaqueous fractionation method has been shown to be appropriate for determining the subcellular distribution of metabolites in spinach leaf cells (Gerhardt and Heldt, 1984). Previously, Yamada et al. (2009a) developed a method for analyzing the subcellular distribution of carbohydrates using the nonaqueous fractionation method in combination with an infiltration-centrifugation method.

In the present study, we identified 2-C-ME as a major soluble carbohydrate in P. drummondii petals. The number and size of petal cells was investigated to confirm whether petal growth during flower opening is mainly dependent on cell expansion. Furthermore, we determined subcellular carbohydrate concentrations in the P. drummondii petals to highlight the significance of soluble carbohydrates, in particular 2-C-ME, as osmotica.

Materials and Methods

Plant material

Seedlings of Phlox drummondii ‘Dolly Sky Blue’ were individually planted in 9-cm diameter plastic pots that were filled with potting medium (Metro-Mix 350; Hyponex Japan, Osaka, Japan) in the middle of autumn. The seedlings were grown in a greenhouse until the next spring under natural day length conditions with temperatures between 15°C and 25°C. Liquid fertilizer (Minelap; Sumitomo Chem., Tokyo, Japan) was applied weekly to the plants. In April or May, the flowers were collected for use in experiments.

Petals were harvested at five different stages of flower development as shown in Figure 1A: petals had not pigmented (stage 1); petals started to pigment (stage 2); petals were pigmented, but the petal tube was not seen (stage 3); flowers started to open, with a petal tube length of 3 mm (stage 4); and flowers within one day of opening (stage 5). Petal limbs were cut from the petals and used for the subsequent experiments.

Fig. 1

Developmental stages (A) and fresh weight of petals (B) during flower opening in P. drummondii. The scale bar represents 1 cm. Values are the means of eight independent experiments ± SE. Different letters indicate significant differences (P < 0.05) by Tukey-Kramer’s multiple range test.

Identification of 2-C-methyl-d-erythritol

Approximately 10 g fresh weight (FW) of petal limbs was cut from open flowers, immersed in 200 mL of 80% ethanol at 75°C for 30 min, and then homogenized. The homogenate was centrifuged at 3,000 × g for 10 min. The supernatant was concentrated in vacuo below 50°C. The concentrate was passed through a Sep-Pak C18 cartridge (Merk Millipore, Milford, MA, USA) to remove hydrophobic compounds. The eluate was concentrated and purified using a high-performance liquid chromatography (HPLC) system (JASCO, Tokyo, Japan) equipped with a refractive-index detector. The sample was applied to a Shodex NH2 column (5E; Showa Denko, Tokyo, Japan), which was eluted with 75% acetonitrile at a flow rate of 3 mL·min⁻¹. After removing the acetonitrile, the fraction containing the unidentified substance was purified on a Shodex SUGAR SP0810 column (Showa Denko), which was maintained at 80°C and eluted with water at a flow rate of 0.8 mL·min⁻¹ to yield ca. 45 mg of syrup. The purified compound was lyophilized and subjected to analysis using NMR. The ¹H-NMR spectra were measured in D₂O at 500 MHz using a JEOL JNM-A500 instrument (Tokyo, Japan). Acetone was used as an internal standard.

Measurement of dry weight and area of petal limbs

The petal limbs were dried at 120°C in an oven for three days, and the dry weight was determined. The area of petal limbs was measured using a leaf area counter (AAM-9; Hayashi Denko, Tokyo, Japan).

Measurements of adaxial and abaxial epidermal cell numbers

Petal limb tissues were fixed in FPA solution (formaldehyde:propionic acid:ethanol = 5:5:90) at 4°C for five days, followed by washing in 50% then 30% ethanol for 30 min each. The tissues were cut into 2-mm squares and immersed in a transparency solution, which was prepared by dissolving 8 g trichloroacetoaldehyde monohydrate in 1 mL glycerol and 2 mL H₂O. The epidermal cells were observed using the Nomarski differential interference contrast method with an optical microscope (AX-70; Olympus, Tokyo, Japan). The number of cells in five fields of view (23,674 μm²), which were randomly selected from one flower, were counted. The total cell number per petal limb was calculated by multiplying the number of cells counted by the total petal limb area. The average area of individual cells was determined by one angle area divided by the cell number.

Carbohydrate extraction and determination

After collecting the petal limbs (five stages), stem, and leaves from plants, soluble carbohydrates were extracted from these organs, and their contents were determined, as described by Yamada et al. (2009a). To measure the starch content, the residue that remained following soluble sugar extraction was dried in vacuo, resuspended in DMSO, and extracted at 100°C for 30 min. The glucose liberated by glucoamylase was measured using the GlucoseTest-Wako (Fujifilm Wako Chemical, Osaka, Japan). Water-soluble pectin was extracted using the method described by Terasaki et al. (2001), and the uronic acid content was estimated using the m-hydroxydiphenyl method (Blumenkrantz and Asboe-Hansen, 1973).

Nonaqueous fractionation method

A nonaqueous fractionation method was used as described in Stitt et al. (1989) with slight modifications. Petal limbs were homogenized with a pre-cooled mortar and pestle in liquid nitrogen and the frozen powder was dried in a lyophilizer for more than four days. The dried powder was suspended in 20 mL of heptane and ultrasonicated for a total 90 sec. The suspension was then screened through 80 μm nylon mesh and centrifuged at 2,000 × g for 10 min. The supernatant was discarded and the sediment was re-suspended in 2 mL of a tetrachloroethylene–heptane mixture (1.28 g·cm⁻³), and shaken well. The mixture was applied to a density gradient comprising a cushion of CCl₄ overlaid with a linear gradient of tetrachloroethylene–heptane mixture decreasing from 1.55 to 1.40 g·cm⁻³. The gradients were centrifuged at 15,000 × g for 15 h, and the contents of the centrifuge tubes were removed from the top and separated into seven fractions. Each fraction was diluted 3-fold with heptane, and centrifuged at 10,000 × g for 10 min, after which the sediments were evaporated to dryness in vacuo for 15 h, and then assayed for markers and carbohydrates.

Collection of apoplastic carbohydrates using an infiltration-centrifugation method

Petal limbs (1 g FW) were cut into 5-mm squares and washed in deionized water to remove soluble proteins and carbohydrates from the cut surfaces. The petal limb pieces were then vacuum-infiltrated in 5 mM MES-NaOH buffer (pH 6.0) for 20 min. The pieces were blotted dry and placed vertically on 0.45-μm filters (Ultrafree CL; Merk Millipore) in the upper compartment of disposable tubes to which 10 mM HEPES-NaOH buffer (pH 8.0) had been added. The tubes were centrifuged at 800 × g for 10 min. The soluble carbohydrate content of the resulting apoplastic fluid was determined by HPLC (JASCO). Recovery of apoplastic fluid was conducted by the method of Yamada et al. (2009a). The eluent was used to measure soluble marker enzyme activity.

Extraction and assay of NADP-glucose-6-phosphate dehydrogenase and anthocyanin

NADP-glucose-6-phosphate dehydrogenase (NADP-G6PDH; EC 1.1.1.49, a cytoplasmic marker) was extracted from the dried sediments prepared using nonaqueous fractionation, and its activity was measured using the methods of Kornberg and Horecker (1955). Anthocyanin (a vacuolar marker) was extracted from the dried segments, and its content was determined as described in Ichimura and Hiraya (1999).

Transmission electron microscopy

Transmission electron micrographs of petal limbs at stage 5 were prepared as previously described by Norikoshi et al. (2015). Forty micrographs were taken at random.

Calculation of subcellular carbohydrate concentrations

Subcellular distribution of soluble carbohydrates was determined by the nonaqueous fractionation method combined with the infiltration-centrifugation method as previously described in Yamada et al. (2009a) with a slight modification. Markers of vacuole, cytoplasm and apoplast were anthocyanin, NADP-G6PDH and soluble pectin, respectively. The fractionation pattern of three markers in one experiment is shown in Supplementary Figure S1. The carbohydrate content in the apoplast fraction was determined by the infiltration-centrifugation method. Apoplastic substances, including carbohydrates, were fractionated by the nonaqueous fractionation method and the distribution pattern could be determined using an apoplastic marker. After fractionation of carbohydrates in a nonaqueous gradient, the estimated apoplastic carbohydrate content was subtracted from each fraction. Then, the carbohydrate contents in the cytoplasm and the vacuole were calculated by the distribution pattern of the two markers using the method of Riens et al. (1991).

To calculate subcellular carbohydrate concentrations, volumes of the vacuole, cytoplasm, cell wall and air space were determined based on their area ratio determined using transmission electron micrographs using the method of Yamada et al. (2009a). Density of petal limbs was calculated based on the volumes of air space and other parts. Subcellular carbohydrate concentrations were calculated by the equation described in Yamada et al. (2009a).

Measurement of inorganic ion concentrations

Petal limb pieces (0.2 g FW) were frozen in liquid nitrogen, placed in a centrifugal filter device (Merk Millipore), and centrifuged at 12,000 × g for 10 min. The resulting fluids were used to determine inorganic ion concentrations by HPLC, as previously described by Norikoshi et al. (2013).

Measurement of osmotic potential

Petal limb pieces (0.2 g FW) were immersed in 2.5 mL of distilled water and heated at 100°C for 20 min. The osmotic potential of the resulting solution was measured using a vapor pressure osmometer (VAPRO5520; WESCOR, Logan, UT, USA), according to the instruction manual.

Statistical analysis

The Student’s t-test and Tukey-Kramer’s multiple range test were performed using SigmaPlot software (v.12.5; Systat Software, San Jose, CA, USA).

Results

Changes in fresh and dry weight during petal development

The fresh weight (FW) of petal limbs and tubes increased during flower opening (Fig. 1B). The FW of petals increased from 11 mg at stage 1 to 41 mg at stage 5.

Figure 2 shows changes in the FW and dry weight (DW) of petal limbs during flower opening. The FW of petal limbs at stage 1 was 6 mg; this gradually increased to almost 32 mg at stage 5 (Fig. 2A). The DW of petal limbs also increased throughout the stages (Fig. 2A). The FW/DW ratio decreased slightly from stages 1 to 4, but increased from stages 4 to 5 (Fig. 2B).

Fig. 2

Fresh and dry weight (A) and the fresh weight/dry weight ratio (B) of petal limbs during flower opening in P. drummondii. Values are the means of five independent experiments ± SE. Different letters indicate significant differences (P < 0.05) by Tukey-Kramer’s multiple range test.

Changes in the number and area of adaxial and abaxial cells

The area of petal limbs increased during flower opening (Fig. 3A). The area of abaxial and adaxial epidermal cells in petal limbs gradually increased during flower opening (Fig. 3B). The number of adaxial epidermal cells was greater than that of abaxial epidermal cells throughout the five stages (Fig. 3C). The number of abaxial and adaxial epidermal cells increased until stage 2 and remained almost constant thereafter.

Fig. 3

Petal limb area (A), epidermal cell area (B) and cell numbers (C) of petal limbs during flower opening in P. drummondii. Values are the means of three independent experiments ± SE. Different letters indicate significant differences (P < 0.05) by Tukey-Kramer’s multiple range test.

Identification of 2-C-methyl-d-erythritol and its concentration in various organs

In addition to glucose, fructose, sucrose, and myo-inositol, a large peak (A) was detected (Supplementary Fig. S2). The retention time of this peak was the same as seen for 2-C-ME isolated from P. subulata (Enomoto et al., 2004). A compound corresponding to this peak was isolated and subjected to ¹H-NMR analysis. The ¹H-NMR spectrum was identical to that of 2-C-ME as shown in Sakamoto et al. (2000). Thus, the isolated compound was identified as 2-C-ME.

2-C-ME was the most abundant carbohydrate in the petal limbs. In the stems and leaves, 2-C-ME was the second most abundant carbohydrate, and the sucrose concentration was much higher than that of the 2-C-ME concentration in these organs (Table 1).

Table 1

Soluble carbohydrate concentrations in petals, stem and leaves.

Soluble carbohydrate concentration and starch content in petal limbs during flower opening

Figure 4 shows the soluble carbohydrate concentration and starch content in the petal limbs during flower opening. 2-C-ME was the most abundant carbohydrate in all five stages, and its concentration markedly increased from stages 4 to 5. Glucose was not detected in stages 1 or 2. However, glucose was the second most abundant carbohydrate from stages 3 to 5, and its concentration increased during flower opening. Sucrose was not detected prior to stage 3. The sucrose concentration peaked at stage 3, but it was present at much lower concentrations than 2-C-ME and glucose. Fructose and myo-inositol concentrations were relatively low and did not increase during flower opening. The starch content at stage 1 was 2.9 mg·g⁻¹FW; this gradually decreased during flower opening.

Fig. 4

Soluble carbohydrate concentrations (A) and starch content (B) in petal limbs during flower opening in P. drummondii. Values are the means of three independent experiments ± SE. Different letters indicate significant differences (P < 0.05) at each stage (A) and among stages (B) by Tukey-Kramer’s multiple range test.

Inorganic ion concentrations in petal limbs during flower opening

In the petal limbs, the cations detected were NH₄⁺, K⁺, Ca²⁺, Mg²⁺, and Na⁺, while the anions detected were NO₃⁻, H₂PO₄⁻, SO₄²⁻, and Cl⁻ (Fig. 5). The concentration of K⁺ was much higher than that of other ions at each of the five stages. The K⁺ concentration was 104 mM at stage 1; this decreased at stage 2 and then increased again between stages 3 and 4. Although the H₂PO₄⁻ concentration was lower than the Cl⁻ concentration at stage 1, the H₂PO₄⁻ concentration increased between stages 2 and 4. The total inorganic ion concentration decreased between stages 1 and 2, and tended to increase thereafter.

Fig. 5

Inorganic ion concentrations in petal limbs during flower opening in P. drummondii. (A) Cation concentration. (B) Anion concentration. Values are the means of three independent experiments ± SE. Different letters indicate significant differences (P < 0.05) at each stage by Tukey-Kramer’s multiple range test.

Osmotic potential in petal limbs during flower opening

The osmotic potential at stages 1 and 2 was determined using two samples because the petal limbs at these stages were very small. However, the differences between the samples were relatively small, suggesting that the data were reliable. The osmotic potential of the petal limbs increased from stages 1 to 3 and decreased thereafter (Fig. 6A). The osmotic potential due to soluble carbohydrates decreased during flower opening (Fig. 6B). The osmotic potential due to inorganic ions increased between stages 1 and 2, but decreased between stages 2 and 4.

Fig. 6

Osmotic potential of petal limbs during flower opening in P. drummondii. (A) Total osmotic potential. Values are the means of two (stages 1 and 2) or four (stages 3, 4, and 5) independent experiments ± SE, and ** is significant at P < 0.05 between stages 4 and 5 by Student’s t-test. (B) Osmotic potential due to carbohydrates and inorganic ions. The values of osmotic potential due to carbohydrates and inorganic ions were calculated from data shown in Figures 4 and 5, respectively. Values are the means of three independent experiments ± SE. Different letters indicate significant differences (P < 0.05) by Tukey-Kramer’s multiple range test.

Subcellular carbohydrate concentration in petal limbs

The subcellular distribution of carbohydrates in the petal limbs was investigated at stage 5. Glucose, fructose, sucrose, myo-inositol, and 2-C-ME were mainly distributed in the vacuole. 2-C-ME and glucose were also distributed in the apoplast, whereas sucrose was distributed in the cytoplasm. The relative ratio for the areas of the vacuole, cell wall, cytoplasm, and air spaces was determined by observing transmission electron micrographs (Fig. 7). The cells of petal limbs were largely occupied by their vacuole. The volumes of the vacuole and the air space were 471 and 56 mm³·g⁻¹ FW, respectively.

Fig. 7

Transmission electron micrograph (A), subcellular volumes (B), and subcellular carbohydrate concentrations (C) of petal limbs at stage 5 in P. drummondii. The scale bar represents 10 μm. Values are the means of three independent experiments ± SE. Different letters indicate significant differences (P < 0.05) among compartments (B) and among each compartment (C) by Tukey-Kramer’s multiple range test.

Soluble carbohydrate concentrations in each compartment in the petal limb cells were calculated from the carbohydrate content and the volume of each compartment. 2-C-ME and glucose concentrations in the vacuole were 131 and 60 mM, respectively. The 2-C-ME concentration in the apoplast was relatively high, but 2-C-ME was not detected in the cytoplasm. Although the sucrose concentration in the vacuole was relatively low (7 mM), its concentration in the cytoplasm was 22 mM, which was the highest concentration of any of the carbohydrates in the cytoplasm.

Discussion

Flower opening is associated with the expansion of petal cells. Consequently, accumulation of osmotica in petal cells is required for cell expansion. In this study, we identified 2-C-ME as the main soluble carbohydrate in P. drummondii petal limbs. 2-C-ME concentrations increased during flower opening, with the concentration of this carbohydrate in the vacuole being much higher than that of any other carbohydrates.

In higher plants, 2-C-ME has been reported in just four plant species (Anthonsen et al., 1976; Dittrich and Angyal, 1988; Ahmed et al., 1996; Enomoto et al., 2004). Although 2-C-ME is a major soluble carbohydrate in P. subulata (Enomoto et al., 2004), its concentrations in the other plant species have not yet been reported. In the present study, we identified 2-C-ME as a major soluble carbohydrate in P. drummondii (Table 1). Thus, plants of the genus Phlox may specifically contain large amounts of this compound.

Petals of P. drummondii consist of limbs and tubes. Measurements of petal limb and tube FW suggest that limbs grew more conspicuously than tubes during flower opening (Fig. 1B). Therefore, we used petal limbs to determine the cell area, soluble carbohydrates, inorganic ions and osmotic potential.

The number of adaxial and abaxial epidermal cells did not increase from stage 2 (Fig. 3), indicating that cell division within the petal limbs ceases almost completely at around this stage. In contrast, the size of the epidermal cells increased during flower opening (Fig. 3). These results indicate that the petal growth associated with flower opening in P. drummondii is due to cell expansion. Similar findings have been reported in rose (Yamada et al., 2009b), E. grandiflorum (Norikoshi et al., 2016), G. grandiflora (Koning, 1984), and T. caerulea (Norikoshi et al., 2013).

For cell expansion to occur, an accumulation of compounds that decreases the osmotic potential in cells, followed by water influx, is required. In the petal limbs, the FW/DW ratio increased (Fig. 2B), as did the soluble carbohydrate concentration, during flower opening (Fig. 4). An increase in the FW/DW ratio during flower opening has also been reported in daylily (Bieleski, 1993), rose (Yamada et al., 2009a), and T. caerulea (Norikoshi et al., 2013). The osmotic potential of the petal limbs decreased between stages 3 and 5, which was associated with increases in soluble carbohydrate concentrations (Fig. 6A). In particular, the total osmotic potential decreased by 123 kPa from stage 4 to stage 5, with a decrease in the osmotic potential of 103 kPa due to soluble carbohydrates. Thus, the decrease in osmotic potential could be mainly attributed to increases in soluble carbohydrate concentrations. Decreases in osmotic potential due to 2-C-ME, glucose, and myo-inositol between stages 4 and 5 were calculated to be 119, 20, and 7 MPa, respectively. Therefore, the contribution of 2-C-ME to osmotic potential was the highest, followed by glucose. However, the osmotic potential due to sucrose and fructose increased between these stages, suggesting that these sugars do not contribute to a decrease in osmotic potential. These results differed from those in daylily and rose, in which glucose and fructose contribute to a decrease in osmotic potential (Bieleski, 1993; Yamada et al., 2009a).

In plants, K⁺ is the primary ion that contributes to the maintenance of osmotic potential, although Ca²⁺, Mg²⁺, NO₃⁻, and Cl⁻ also play a major role in the shoots and leaves of some plant species (Cram, 1976; Leigh and Tomos, 1993). In the petal limbs of P. drummondii, the concentration of K⁺ was highest among the inorganic ions, while H₂SO₄²⁻ and Cl⁻ concentrations were also relatively high (Fig. 5). During stages 1 and 2, the osmotic potential due to inorganic ions was much lower than the osmotic potential due to soluble carbohydrates, suggesting that inorganic ions are important osmotica during the young stages. However, the osmotic potential due to inorganic ions did not decrease between stages 4 and 5. Thus, the role played by inorganic ions as osmotica during flower opening appears to be limited.

The total osmotic potential was lower than the sum of the osmotic potential due to soluble carbohydrates and inorganic ions, suggesting that other compounds are involved in maintaining osmotic potential. Since the total cation concentration was much higher than the total anion concentration, anionic compounds, such as organic acids, are likely to be osmotica. Organic acids have been shown to act as osmotica in many plants, such as in sugarcane stalk (Cram, 1976; Welbaum and Meinzer, 1990). In addition to these compounds, amino acids and polyphenol compounds are also suggested to be involved in the maintenance of osmotic potential (Yamaki, 1984; Ichimura et al., 2016). These compounds may be osmotica in P. drummondii petals.

The concentration of 2-C-ME in petal limbs was much higher than the concentrations of the other soluble carbohydrates at each of the five stages, and this increased during flower opening (Fig. 4). For cell expansion, some compounds, including soluble carbohydrates, amino acids, and inorganic ions, accumulate in the vacuole as osmotica and contribute to cell expansion (Yamaki, 1984; Yamada et al., 2009a). As a result, we investigated subcellular concentrations of soluble carbohydrates in the petal limb. The concentration of 2-C-ME in the vacuole of open flowers was estimated to be 131 mM, which was much higher than the concentrations of other carbohydrates (Fig. 7). This suggests that 2-C-ME is the most important carbohydrate for maintaining the osmotic potential for petal cell expansion. For cell expansion, substrates for respiration and cell wall synthesis are required. In general, glucose, fructose, and sucrose, which are primarily metabolic sugars, are known to act as the substrates for respiration. Polyol has been shown to have a role as a reserve carbohydrate (Loescher, 1987). Since 2-C-ME belongs to polyol, 2-C-ME may be a reserve carbohydrate of which the metabolite acts as a substrate for respiration.

The concentration of 2-C-ME was relatively high in the apoplast (Fig. 7). Translocated sugars are unloaded through the apoplast and/or symplast into sink organs (Patrick and Offler, 1996). In tomato (Damon et al., 1988) and apple fruits (Zhang et al., 2004), in which sugar concentrations in the apoplast are relatively high, sugars are at least partially transported through the apoplast. Some polyols, including mannitol and sorbitol, are known to be carbohydrates used for translocation (Loescher, 1987). Therefore, we can hypothesize that 2-C-ME is a carbohydrate used for translocation that accumulates in the apoplast through apoplastic unloading in P. drummondii.

Although glucose was not detected during stages 1 or 2, its concentration was the second highest of all the carbohydrates at each of the other stages (Fig. 4). In addition, the glucose concentration was the second highest in the vacuole at stage 5 (Fig. 7), suggesting that glucose is also an important carbohydrate that acts as an osmoticum. Glucose is a major osmoticum in the petals of many plants including rose (Yamada et al., 2009a), chrysanthemum (Ichimura et al., 2000b) and carnation (Ichimura et al., 1998).

The concentration of sucrose in petal limbs was lower than that of glucose and 2-C-ME, at both stages 4 and 5 (Fig. 4). Although sucrose was distributed in the vacuole, the sucrose concentration in the cytoplasm was relatively high (Fig. 7). This result is consistent with findings seen for Hippeastrum and tulip (Wagner, 1979) and rose (Yamada et al., 2009a). Sucrose is known to be generally synthesized in the cytoplasm (Bird et al., 1974), whereas the activities of acid invertase localized in the vacuole and insoluble invertase localized in the apoplast are relatively high (Yamada et al., 2007). The accumulation of sucrose in the cytoplasm may be attributable to the activities of these enzymes.

In the petals of many plants, including chrysanthemum (Ichimura et al., 2000b), rose (Yamada et al., 2009a), carnation (Ichimura et al., 1998), gladiolus (Yamane et al., 1991), Tweedia caerulea (Norikoshi et al., 2013), glucose and fructose accumulate in the petals during flower opening. In addition to glucose, sucrose, rather than fructose, accumulates in the petals of sweet pea (Ichimura et al., 1999) and E. grandiflorum (Norikoshi et al., 2016). However, the mechanisms underlying sugar accumulation have not been studied. In P. drummondii petals, the 2-C-ME concentration was much higher than the other carbohydrate concentrations, whereas accumulation of fructose was relatively low. Based on the experiments using Liriodendron tulipifera leaves, Sagner et al. (1998) proposed that 2-C-ME is produced from glyceraldehyde 3-phosphate (G3P) through a non-mevalonate pathway via several enzymatic reactions (Fig. 8). Fructose in the petal limbs only slightly accumulated during flower opening (Fig. 4), although G3P is produced from sucrose through fructose. In higher plants, fructose is metabolized to fructose 1-phosphate (F1P) by fructokinase and F1P is metabolized to G3P (Kanayama et al., 1998). As shown in Figure 8, fructokinase is involved in the synthesis of 2-C-ME and fructose metabolism. Therefore, we propose that the activity of fructokinase and enzymes involved in 2-C-ME synthesis is high in P. drummondii, resulting in the large accumulation of 2-C-ME and slight accumulation of fructose. 2-C-Methyl-d-erythritol 4-phosphate has been shown to be a precursor of abscisic acid (ABA) (Milborrow, 2001). Flower opening in Japanese morning glory (Kaihara and Takimoto, 1983) and wintersweet (Sui et al., 2015) is promoted by exogenous ABA. Therefore, it can be hypothesized that ABA metabolized from accumulated 2-C-ME contributes to flower opening in P. drummondii.

Fig. 8

Proposed pathway for 2-C-methyl-d-erythritol based on Sagner et al. (1998).

Mechanisms underlying flower opening are still not well understood. In Japanese morning glory, flower opening is caused mainly by the epinasty of the petal midrib (Kaihara and Takimoto, 1981). Similarly, the midrib is important for flower opening in Lilium (Bieleski et al., 2000). In Lilium ‘Casabranca’, however, the removal of the midrib did not suppress flower opening, suggesting that it is not necessary in driving flower opening (Liang and Mahadevan, 2011). Recently, Cheng et al. (2021) reported that petal movement in rose flowers is driven by the asymmetric growth of the petal base, primarily due to an increase in the parenchyma cell size in the adaxial side of the petal. P. drummondii petals consists of petal limbs and tubes (Fig. 1). The petal tubes in P. drummondii seem to correspond to the base of petals in rose. In our study, petal limbs were used as a material because marked growth was observed during flower opening. To clarify the mechanism underlying flower opening in P. drummondii, it may be necessary to study whether the petal tubes have a role similar to that of the petal base in rose.

The total carbohydrate concentration in the petals is assumed to decrease during flower opening because cell expansion is known to be driven by water influx. However, in the present study, soluble carbohydrate concentrations reached their highest at stage 5 (Fig. 4). In addition, the osmotic potential due to soluble carbohydrates was much lower in the symplast than apoplast at stage 5 (Fig. 7). In many plants, including Delphinium (Ichimura et al., 2000a) and E. grandiflorum (Kawabata et al., 2011), petal growth continues after flower opening. Thus, the concentration of soluble carbohydrates in the petal limbs may decrease after flower opening. Alternatively, the termination of cell expansion may be due to decreases in cell wall extensibility, because cell expansion requires cell wall extensibility as well as water influx.

In conclusion, we identified 2-C-ME as a major soluble carbohydrate in P. drummondii petals. During flower opening, the concentration of 2-C-ME in petal limbs increased. In the petals of open flowers, the concentration of 2-C-ME in the vacuole was higher than 100 mM. The osmotic potential decreased with flower development, and this decrease was mainly attributed to an increase in the 2-C-ME concentration. Therefore, the accumulation of 2-C-ME at high concentrations in the vacuole of petal limbs appears to contribute to the cell expansion associated with flower opening.

Acknowledgements

We would like to thank Dr. M. Nakayama of NARO for his advice on the biosynthetic pathway of 2-C-ME.

Literature Cited

•  Ahmed,  A. A.,  M. H. A.  El-Razek,  E. A. A.  Mostafa,  H. J.  Williams,  A. I.  Scott,  J. H.  Reibenspies and  T. J.  Mabry. 1996. A new derivative of glucose and 2-C-methyl-d-erythritol from Ferula sinaica. J. Nat. Prod. 59: 1171–1173.
•  Anthonsen,  T.,  S.  Sagen,  M. A.  Kazi,  S. W.  Shah and  S.  Tagar. 1976. 2-C-methyl-d-erythritol, a new branched alditol from Convolvulus glomeratus. Acta Chem. Scand. B30: 91–93.
•  Beauzamy,  L.,  N.  Nakayama and  A.  Boudaoud. 2014. Flowers under pressure: ins and outs of turgor regulation in development. Ann. Bot. 114: 1517–1533.
•  Bieleski,  R. L. 1993. Fructan hydrolysis drives petal expansion in the ephemeral daylily flower. Plant Physiol. 103: 213–219.
•  Bieleski,  R. L.,  H. J.  Elgar and  J. A.  Heyes. 2000. Mechanical aspects of rapid flower opening in Asiatic lily. Ann. Bot. 86: 1175–1183.
•  Bird,  I. F.,  M. J.  Cornelius,  A. J.  Keys and  C. P.  Whittingham. 1974. Intracellular site of sucrose synthesis in leaves. Phytochemistry 13: 59–64.
•  Blumenkrantz,  N. and  G.  Asboe-Hansen. 1973. New method for quantitative determination of uronic acids. Analytical Biochem. 54: 484–489.
•  Cheng,  C.,  Q.  Yu,  Y.  Wang,  H.  Wang,  Y.  Dong,  Y.  Ji,  X.  Zhou,  Y.  Li,  C.-Z.  Jiang,  S.-S.  Gan,  L.  Zhao,  Z.  Fei,  J.  Gao and  N.  Ma. 2021. Ethylene-regulated asymmetric growth of the petal base promotes flower opening in rose (Rosa hybrida). Plant Cell 33: 1229–1251.
• Cram, W. J. 1976. Negative feedback regulation of transport in cells. The maintenance of turgor, volume and nutrient supply. p. 284–316. In: U. Lüttge and M. G. Pitman (eds.). Encyclopedia of Plant Physiology. New Series Vol. 2. Springer-Verlag, Berlin.
•  Damon,  S.,  J.  Hewitt,  M.  Nieder and  A.  Bennett. 1988. Sink metabolism in tomato fruit. II. Phloem unloading and sugar uptake. Plant Physiol. 87: 731–736.
•  Dittrich,  P. and  S. J.  Angyal. 1988. 2-C-methyl-d-erythritol in leaves of Liriodendron tulipifera. Phytochemisry 27: 935.
•  Enomoto,  H.,  K.  Kohata,  M.  Nakayama,  Y.  Yamaguchi and  K.  Ichimura. 2004. 2-C-methyl-d-erythritol is a major carbohydrate in petals of Phlox subulata possibly involved in flower development. J. Plant Physiol. 161: 977–980.
•  Gerhardt,  R. and  H. W.  Heldt. 1984. Measurement of subcellular metabolite levels in leaves by fractionation of freeze-stopped material in nonaqueous media. Plant Physiol. 75: 542–547.
•  Ichimura,  K. and  T.  Hiraya. 1999. Effect of silver thiosulfate complex (STS) in combination with sucrose on the vase life of cut sweet pea flowers. J. Japan. Soc. Hort. Sci. 68: 23–27.
•  Ichimura,  K.,  Y.  Kawabata,  M.  Kishimoto,  R.  Goto and  K.  Yamada. 2003. Shortage of soluble carbohydrates is largely responsible for short vase life of cut ‘Sonia’ rose flowers. J. Japan. Soc. Hort. Sci. 72: 292–298.
•  Ichimura,  K.,  K.  Kohata and  R.  Goto. 2000a. Soluble carbohydrates in Delphinium and their influence on sepal abscission in cut flowers. Physiol. Plant. 108: 307–313.
•  Ichimura,  K.,  K.  Kohata,  M.  Koketsu,  M.  Shimamura and  A.  Ito. 1998. Identification of pinitol as a main sugar constituent and changes in its content during flower bud development in carnation (Dianthus caryophyllus L.). J. Plant Physiol. 152: 363–367.
•  Ichimura,  K.,  K.  Kohata,  Y.  Mukasa,  Y.  Yamaguchi,  R.  Goto and  K.  Suto. 1999. Identification of l-bornesitol and changes in its content during flower bud development in sweet pea (Lathyrus odoratus L.). Biosci. Biotechnol. Biochem. 63: 189–191.
•  Ichimura,  K.,  K.  Kohata,  Y.  Yamaguchi,  M.  Douzono,  H.  Ikeda and  M.  Koketsu. 2000b. Identification of l-inositol and scyllytol and their distribution in various organs in chrysanthemum. Biosci. Biotechnol. Biochem. 64: 865–868.
•  Ichimura,  K.,  H.  Ono,  A.  Soga,  H.  Shimizu-Yumoto,  K.  Kohata and  M.  Nakayama. 2016. 2-Cyanoethyl-isoxazolin-5-one is a major low molecular weight nitrogenous compound in sweet pea (Lathyrus odoratus L.). J. Japan. Soc. Hort. Sci. 85: 161–168.
•  Kaihara,  S. and  A.  Takimoto. 1981. Physical basis of flower opening in Pharbitis nil. Plant Cell Physiol. 22: 307–310.
•  Kaihara,  S. and  A.  Takimoto. 1983. Effect of plant growth regulators on flower-opening of Pharbitis nil. Plant Cell Physiol. 24: 309–316.
•  Kanayama,  Y.,  D.  Granot,  N.  Dai,  M.  Petreikov,  A.  Schaffer,  A.  Powell and  A. B.  Bennett. 1998. Tomato fructokinases exhibit differential expression and substrate regulation. Plant Physiol. 117: 85–90.
•  Kawabata,  S.,  K.  Nii and  M.  Yokoo. 2011. Three-dimensional formation of corolla shapes in relation to the developmental distortion of petals in Eustoma grandiflorum. Sci. Hortic. 132: 66–70.
•  Keller,  M. and  P. M.  Shrestha. 2014. Solute accumulation differs in the vacuoles and apoplast of ripening grape berries. Planta 239: 633–642.
•  Koning,  R. E. 1984. The role of plant hormones in the growth of the corolla of Gaillardia grandiflora (Asteraceae) ray flowers. Amer. J. Bot. 71: 1–8.
•  Kornberg,  A. and  B. L.  Horecker. 1955. Glucose-6-phosphate dehydrogenase. Methods Enzymol. 1: 323–325.
•  Leigh,  R. A. and  A. D.  Tomos. 1993. Ion distribution in cereal leaves: Pathways and mechanisms. Phil. Trans. R. Soc. Lond. B 341: 75–86.
•  Liang,  H. and  L.  Mahadevan. 2011. Growth, geometry, and mechanics of a blooming lily. Proc. Natl. Acad. Sci. USA 108: 1302–1305.
•  Loescher,  W. H. 1987. Physiology and metabolism of sugar alcohols in higher plants. Physiol. Plant. 70: 553–557.
•  Milborrow,  B. V. 2001. The pathway of biosynthesis of abscisic acid in vascular plants: a review of the present state of knowledge of ABA biosynthesis. J. Exp. Bot. 52: 1145–1164.
•  Norikoshi,  R.,  H.  Imanishi and  K.  Ichimura. 2013. Changes in cell number, osmotic potential and concentrations of carbohydrates and inorganic ions in Tweedia caerulea during flower opening. J. Japan. Soc. Hort. Sci. 82: 51–56.
•  Norikoshi,  R.,  K.  Kohata,  H.  Shimizu-Yumoto,  R.  Goto and  K.  Ichimura. 2016. Identification of soluble carbohydrates and their subcellular concentrations in petals during flower opening in Eustoma grandiflorum. Hort. J. 85: 238–247.
•  Norikoshi,  R.,  K.  Yamada,  T.  Niki and  K.  Ichimura. 2015. Accumulation of mannitol in the cytoplasm and vacuole during the expansion of sepal cells associated with flower opening in Delphinium × belladonna cv. Bellamosum. Planta 242: 1467–1477.
•  Ofosu-Anim,  J. and  S.  Yamaki. 1994. Sugar content and compartmentation in melon fruit and the restriction of sugar efflux from flesh tissue by ABA. J. Japan. Soc. Hort. Sci. 63: 685–692.
•  Patrick,  J. W. and  C. E.  Offler. 1996. Post-sieve element transport of photoassimilates in sink regions. J. Exp. Bot. 47: 1165–1177.
•  Paul,  M. and  W.  Cockburn. 1989. Pinitol, a compatible solute in Mesembryanthemum crystallinum L.? J. Exp. Bot. 40: 1093–1098.
•  Riens,  B.,  G.  Lohaus,  D.  Heineke and  H. W.  Heldt. 1991. Amino acid and sucrose content determined in the cytosolic, chloroplastic, and vacuolar compartments and in the phloem sap of spinach leaves. Plant Physiol. 97: 227–233.
•  Sagner,  S.,  W.  Eisenreich,  M.  Fellermeier,  C.  Latzel,  A.  Bacher and  M. H.  Zenk. 1998. Biosynthesis of 2-C-methyl-d-erythritol in plants by rearrangement of the terpenoid precursor, 1-deoxy-d-xylulose 5-phosphate. Tetrahedron Lett. 39: 2091–2094.
•  Sakamoto,  I.,  K.  Ichimura and  H.  Ohrui. 2000. Synthesis of 2-C-methyl-d-erythritol and 2-C-methyl-l-threitol; determination of the absolute configuration of 2-C-methyl-1,2,3,4-butanetetrol isolated from Phlox sublata L. Biosci. Biotechnol. Biochem. 64: 1915–1922.
•  Stitt,  M.,  R. M.  Lilley,  R.  Gerhardt and  H. W.  Heldt. 1989. Metabolite levels in specific cells and subcellular compartments of plant leaves. Methods Enzymol. 174: 518–550.
•  Sui,  S.,  J.  Luo,  D.  Liu,  J.  Ma,  W.  Men,  L.  Fan,  Y.  Bai and  M.  Li. 2015. Effects of hormone treatments on cut flower opening and senescence in wintersweet (Chimonanthus praecox). HortScience 50: 1365–1369.
•  Terasaki,  S.,  N.  Sakurai,  R.  Yamamoto,  N.  Wada and  D. J.  Nevins. 2001. Changes in cell wall polysaccharides of kiwifruit and the visco-elastic properties detected by a laser Doppler method. J. Japan. Soc. Hort. Sci. 70: 572–580.
•  Wagner,  G. J. 1979. Content and vacuole/extravacuole distribution of neutral sugars, free amino acids, and anthocyanin in protoplasts. Plant Physiol. 64: 88–93.
•  Welbaum,  G. E. and  F. C.  Meinzer. 1990. Compartmentation of solutes and water in developing sugarcane stalk tissue. Plant Physiol. 93: 1147–1153.
•  Yamada,  K.,  M.  Ito,  T.  Oyama,  M.  Nakada,  M.  Maesaka and  S.  Yamaki. 2007. Analysis of sucrose metabolism during petal growth of cut roses. Postharvest Biol. Technol. 43: 174–177.
•  Yamada,  K.,  R.  Norikoshi,  K.  Suzuki,  H.  Imanishi and  K.  Ichimura. 2009a. Determination of subcellular concentrations of soluble carbohydrates in rose petals during opening by nonaqueous fractionation method combined with infiltration-centrifugation method. Planta 230: 1115–1127.
•  Yamada,  K.,  R.  Norikoshi,  K.  Suzuki,  T.  Nishijima,  H.  Imanishi and  K.  Ichimura. 2009b. Cell division and expansion growth during rose petal development. J. Japan. Soc. Hort. Sci. 78: 356–362.
•  Yamaki,  S. 1984. Isolation of vacuoles from immatured apple fruit flesh and compartmentation of sugars, organic acids, phenolic compounds and amino acids. Plant Cell Physiol. 25: 151–166.
•  Yamaki,  S. and  M.  Ino. 1992. Alteration of cellular compartmentation and membrane permeability to sugars in immature and mature apple fruit. J. Amer. Soc. Hort. Sci. 117: 951–954.
•  Yamane,  K.,  S.  Kawabata and  R.  Sakiyama. 1991. Changes in water relations, carbohydrate contents, and acid invertase activity associated with perianth elongation during anthesis of cut gladiolus flowers. J. Japan. Soc. Hort. Sci. 60: 421–428.
•  Zhang,  L. Y.,  Y. B.  Peng,  S.  Pelleschi-Travier,  Y.  Fan,  Y. F.  Lu,  Y. M.  Lu,  X. P.  Gao,  Y. Y.  Shen,  S.  Delrot and  D. P.  Zhang. 2004. Evidence for apoplastic phloem unloading in developing apple fruit. Plant Physiol. 135: 574–586.